from sklearn_benchmarks.reporting.hp_match import HpMatchReporting
import pandas as pd
pd.set_option('display.max_colwidth', None)
pd.set_option('display.max_columns', None)
pd.set_option('display.max_rows', None)

scikit-learn vs. ONNX Runtime

reporting = HpMatchReporting(other_library="onnx", config="config.yml", log_scale=True)
reporting.make_report()

We assume here there is a perfect match between the hyperparameters of both librairies. For a given set of parameters and a given dataset, we compute the speedup time scikit-learn / time onnx. For instance, a speedup of 2 means that onnx is twice as fast as scikit-learn for a given set of parameters and a given dataset.

KNeighborsClassifier_brute_force

onnx (1.10.1) vs. scikit-learn (1.0.dev0)

Speedup barplots

All estimators share the following parameters: algorithm=brute.

Raw results

predict

	estimator	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_jobs	n_neighbors	accuracy_score_sklearn	mean_duration_onnx	std_duration_onnx	accuracy_score_onnx	speedup	std_speedup	sklearn_profiling	onnx_profiling
0	KNeighborsClassifier_brute_force	predict	100000	1000	100	1.945	0.174	0.000	0.002	-1	1	0.663	17.257	0.031	0.663	0.113	0.113	See	See
1	KNeighborsClassifier_brute_force	predict	100000	1	100	0.023	0.003	0.000	0.023	-1	1	1.000	0.312	0.008	1.000	0.073	0.073	See	See
2	KNeighborsClassifier_brute_force	predict	100000	1000	100	2.759	0.023	0.000	0.003	-1	5	0.757	17.021	0.144	0.757	0.162	0.162	See	See
3	KNeighborsClassifier_brute_force	predict	100000	1	100	0.023	0.003	0.000	0.023	-1	5	1.000	0.315	0.006	1.000	0.074	0.074	See	See
4	KNeighborsClassifier_brute_force	predict	100000	1000	100	2.061	0.034	0.000	0.002	1	100	0.882	16.867	0.023	0.882	0.122	0.122	See	See
5	KNeighborsClassifier_brute_force	predict	100000	1	100	0.034	0.014	0.000	0.034	1	100	1.000	0.333	0.015	1.000	0.101	0.101	See	See
6	KNeighborsClassifier_brute_force	predict	100000	1000	100	2.787	0.043	0.000	0.003	-1	100	0.882	16.957	0.035	0.882	0.164	0.164	See	See
7	KNeighborsClassifier_brute_force	predict	100000	1	100	0.023	0.002	0.000	0.023	-1	100	1.000	0.316	0.005	1.000	0.073	0.073	See	See
8	KNeighborsClassifier_brute_force	predict	100000	1000	100	2.021	0.007	0.000	0.002	1	5	0.757	16.939	0.051	0.757	0.119	0.119	See	See
9	KNeighborsClassifier_brute_force	predict	100000	1	100	0.019	0.000	0.000	0.019	1	5	1.000	0.316	0.006	1.000	0.061	0.061	See	See
10	KNeighborsClassifier_brute_force	predict	100000	1000	100	1.161	0.003	0.001	0.001	1	1	0.663	17.220	0.039	0.663	0.067	0.067	See	See
11	KNeighborsClassifier_brute_force	predict	100000	1	100	0.019	0.002	0.000	0.019	1	1	1.000	0.317	0.006	1.000	0.061	0.061	See	See
12	KNeighborsClassifier_brute_force	predict	100000	1000	2	1.753	0.024	0.000	0.002	-1	1	0.896	3.737	0.009	0.896	0.469	0.469	See	See
13	KNeighborsClassifier_brute_force	predict	100000	1	2	0.006	0.004	0.000	0.006	-1	1	1.000	0.239	0.005	1.000	0.024	0.024	See	See
14	KNeighborsClassifier_brute_force	predict	100000	1000	2	2.641	0.015	0.000	0.003	-1	5	0.922	3.704	0.012	0.922	0.713	0.713	See	See
15	KNeighborsClassifier_brute_force	predict	100000	1	2	0.007	0.005	0.000	0.007	-1	5	1.000	0.240	0.004	1.000	0.030	0.030	See	See
16	KNeighborsClassifier_brute_force	predict	100000	1000	2	1.954	0.002	0.000	0.002	1	100	0.929	3.739	0.007	0.929	0.523	0.523	See	See
17	KNeighborsClassifier_brute_force	predict	100000	1	2	0.003	0.000	0.000	0.003	1	100	1.000	0.240	0.005	1.000	0.012	0.012	See	See
18	KNeighborsClassifier_brute_force	predict	100000	1000	2	2.652	0.023	0.000	0.003	-1	100	0.929	3.742	0.005	0.929	0.709	0.709	See	See
19	KNeighborsClassifier_brute_force	predict	100000	1	2	0.008	0.002	0.000	0.008	-1	100	1.000	0.237	0.005	1.000	0.033	0.033	See	See
20	KNeighborsClassifier_brute_force	predict	100000	1000	2	1.933	0.004	0.000	0.002	1	5	0.922	3.695	0.054	0.922	0.523	0.523	See	See
21	KNeighborsClassifier_brute_force	predict	100000	1	2	0.003	0.000	0.000	0.003	1	5	1.000	0.239	0.005	1.000	0.012	0.012	See	See
22	KNeighborsClassifier_brute_force	predict	100000	1000	2	1.060	0.008	0.000	0.001	1	1	0.896	3.748	0.049	0.896	0.283	0.283	See	See
23	KNeighborsClassifier_brute_force	predict	100000	1	2	0.002	0.000	0.000	0.002	1	1	1.000	0.238	0.004	1.000	0.008	0.008	See	See

KNeighborsClassifier_kd_tree

onnx (1.10.1) vs. scikit-learn (1.0.dev0)

Speedup barplots

All estimators share the following parameters: algorithm=kd_tree.

Raw results

predict

	estimator	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	iteration_throughput	latency	n_jobs	n_neighbors	accuracy_score_sklearn	mean_duration_onnx	std_duration_onnx	accuracy_score_onnx	speedup	std_speedup	sklearn_profiling	onnx_profiling
0	KNeighborsClassifier_kd_tree	predict	1000000	1000	10	0.847	1.024	0.000	0.001	-1	1	0.929	118.190	0.000	0.929	0.007	0.007	See	See
1	KNeighborsClassifier_kd_tree	predict	1000000	1	10	0.003	0.000	0.000	0.003	-1	1	1.000	2.630	0.229	1.000	0.001	0.001	See	See
2	KNeighborsClassifier_kd_tree	predict	1000000	1000	10	1.084	0.368	0.000	0.001	-1	5	0.946	117.113	0.000	0.946	0.009	0.009	See	See
3	KNeighborsClassifier_kd_tree	predict	1000000	1	10	0.003	0.001	0.000	0.003	-1	5	1.000	2.621	0.198	1.000	0.001	0.001	See	See
4	KNeighborsClassifier_kd_tree	predict	1000000	1000	10	5.655	0.496	0.000	0.006	1	100	0.951	115.626	0.000	0.951	0.049	0.049	See	See
5	KNeighborsClassifier_kd_tree	predict	1000000	1	10	0.004	0.001	0.000	0.004	1	100	1.000	2.647	0.197	1.000	0.001	0.001	See	See
6	KNeighborsClassifier_kd_tree	predict	1000000	1000	10	3.316	0.347	0.000	0.003	-1	100	0.951	113.568	0.000	0.951	0.029	0.029	See	See
7	KNeighborsClassifier_kd_tree	predict	1000000	1	10	0.006	0.001	0.000	0.006	-1	100	1.000	2.672	0.233	1.000	0.002	0.002	See	See
8	KNeighborsClassifier_kd_tree	predict	1000000	1000	10	1.803	0.412	0.000	0.002	1	5	0.946	112.714	0.000	0.946	0.016	0.016	See	See
9	KNeighborsClassifier_kd_tree	predict	1000000	1	10	0.002	0.000	0.000	0.002	1	5	1.000	2.627	0.147	1.000	0.001	0.001	See	See
10	KNeighborsClassifier_kd_tree	predict	1000000	1000	10	0.928	0.170	0.000	0.001	1	1	0.929	114.256	0.000	0.929	0.008	0.008	See	See
11	KNeighborsClassifier_kd_tree	predict	1000000	1	10	0.001	0.000	0.000	0.001	1	1	1.000	2.621	0.135	1.000	0.000	0.000	See	See
12	KNeighborsClassifier_kd_tree	predict	1000	1000	2	0.026	0.013	0.001	0.000	-1	1	0.891	0.045	0.015	0.891	0.580	0.613	See	See
13	KNeighborsClassifier_kd_tree	predict	1000	1	2	0.002	0.000	0.000	0.002	-1	1	1.000	0.005	0.000	1.000	0.464	0.464	See	See
14	KNeighborsClassifier_kd_tree	predict	1000	1000	2	0.024	0.001	0.001	0.000	-1	5	0.911	0.045	0.017	0.911	0.527	0.565	See	See
15	KNeighborsClassifier_kd_tree	predict	1000	1	2	0.002	0.000	0.000	0.002	-1	5	1.000	0.005	0.000	1.000	0.450	0.451	See	See
16	KNeighborsClassifier_kd_tree	predict	1000	1000	2	0.035	0.005	0.000	0.000	1	100	0.894	0.070	0.007	0.894	0.500	0.503	See	See
17	KNeighborsClassifier_kd_tree	predict	1000	1	2	0.001	0.000	0.000	0.001	1	100	1.000	0.005	0.000	1.000	0.112	0.112	See	See
18	KNeighborsClassifier_kd_tree	predict	1000	1000	2	0.038	0.004	0.000	0.000	-1	100	0.894	0.061	0.001	0.894	0.623	0.623	See	See
19	KNeighborsClassifier_kd_tree	predict	1000	1	2	0.002	0.000	0.000	0.002	-1	100	1.000	0.005	0.001	1.000	0.415	0.419	See	See
20	KNeighborsClassifier_kd_tree	predict	1000	1000	2	0.020	0.000	0.001	0.000	1	5	0.911	0.039	0.001	0.911	0.514	0.514	See	See
21	KNeighborsClassifier_kd_tree	predict	1000	1	2	0.001	0.000	0.000	0.001	1	5	1.000	0.005	0.000	1.000	0.109	0.109	See	See
22	KNeighborsClassifier_kd_tree	predict	1000	1000	2	0.019	0.000	0.001	0.000	1	1	0.891	0.039	0.000	0.891	0.485	0.485	See	See
23	KNeighborsClassifier_kd_tree	predict	1000	1	2	0.001	0.000	0.000	0.001	1	1	1.000	0.005	0.000	1.000	0.110	0.110	See	See

HistGradientBoostingClassifier_best

onnx (1.10.1) vs. scikit-learn (1.0.dev0)

Speedup barplots

All estimators share the following parameters: learning_rate=0.01, n_iter_no_change=10.0, max_leaf_nodes=100.0, max_bins=255.0, min_samples_leaf=100.0, max_iter=300.0.

Raw results

predict

	estimator	function	n_samples_train	n_samples	n_features	mean_duration_sklearn	std_duration_sklearn	n_iter	iteration_throughput	latency	accuracy_score_sklearn	mean_duration_onnx	std_duration_onnx	accuracy_score_onnx	speedup	std_speedup	sklearn_profiling	onnx_profiling
0	HistGradientBoostingClassifier_best	predict	100000	1000	100	0.108	0.001	300	0.007	0.000	0.824	0.509	0.022	0.824	0.212	0.212	See	See
1	HistGradientBoostingClassifier_best	predict	100000	1	100	0.017	0.000	300	0.000	0.017	1.000	0.421	0.009	1.000	0.040	0.040	See	See